These are chat archives for nightscout/intend-to-bolus

31st
Jul 2017
Marco
@CaptainBalou
Jul 31 2017 00:01
@tim2000s Oh I am sorry - that's what I was thinking about. Mistyped. Thanks for the correction. 👍🏻
Scott Leibrand
@scottleibrand
Jul 31 2017 01:21
@902Lund the docs have directions for upgrading to oref0 dev. That is probably easier than reflashing.
Steve Lund
@902Lund
Jul 31 2017 01:50
Thanks Scott!
drmartinezs
@drmartinezs
Jul 31 2017 01:59
please help, configurating my first rig, in nightscout i have no any data displaying
using dexcom app receiving in galaxy s7
Steve Lund
@902Lund
Jul 31 2017 02:29
Thanks again all! Re-flashed with older Jubilinux 0.1.1 and used the oref0 master and everything looks great in NS now. Having access to this system is truly life altering for T1D's
@drmartinezs I just found out the hard way that if you are using the latest version of Jubilinux on your Edison
@drmartinezs sorry hit enter to soon...if using version 0.2.0 of Jubilnux, you'll have issues with oref0 master. I had problems with NS too.
drmartinezs
@drmartinezs
Jul 31 2017 02:32
oh!
Steve Lund
@902Lund
Jul 31 2017 02:33
If you do have the latest version of Jubilinux, apparently you can use oref0 dev and it should work. I ended up re-flashing with 0.1.1 and everything just lit up for me.
drmartinezs
@drmartinezs
Jul 31 2017 02:34
ok
Scott Leibrand
@scottleibrand
Jul 31 2017 02:39
Someone want to PR in a note to the docs about this?
Steve Lund
@902Lund
Jul 31 2017 02:49
I'll give the PR a shot. It's the least I can do.
Scott Leibrand
@scottleibrand
Jul 31 2017 03:01
Thanks. Merged.
Jake Punshon
@JakePunshon
Jul 31 2017 03:43
I'm getting an A33 error on my Medtronic 715 pump when changing reservoirs. Is there a fix for this, or should I start looking for another pump?
Scott Leibrand
@scottleibrand
Jul 31 2017 03:59
Is the cap at the bottom of the pump motor sticking out, or is it fully depressed like normal? (If it's sticking out, don't try to press it back in without disconnecting from any pump site first, or you might administer way too much insulin.)
Jake Punshon
@JakePunshon
Jul 31 2017 04:04
not sticking out
marafie
@marafie
Jul 31 2017 08:57
I installed the dev branch and I got this
penaps 0.2.0-dev
fatal: destination path 'oref0' already exists and is not an empty directory.
Switched to branch 'master'
Your branch is up-to-date with 'origin/master'.
do I need to empty my oref 1st ?
marafie
@marafie
Jul 31 2017 09:27
fatal: destination path 'oref0' already exists and is not an empty directory.
Already on 'master'
Your branch is up-to-date with 'origin/master'.
fatal: unable to connect to github.com:
github.com[0: 192.30.253.113]: errno=Connection timed out
github.com[1: 192.30.253.112]: errno=Connection timed out
when I ping 192.30.253.112/113 it works
fidoman
@efidoman
Jul 31 2017 11:27
It looks like you can't restart a g5 after replacing the batteries because of the software timeout built-in. Anyone know if that is true or not? See this link where someone tried with a g5 ... they got the batteries replaced but couldn't get the transmitter to restart. http://diabetesemancipation.blogsport.eu/2016/07/16/first-restart-trial-failed/
LilDucky
@LilDucky
Jul 31 2017 11:32
@marafie does your internet firewall/router block any outbound traffic? I had similar issues and had to unblock ssh and other stuff to get the fit clone git://... to work
Andy Sharrow
@dramageek
Jul 31 2017 12:10
Is it documented anywhere what the secret sauce is for using both xdrip as a CGM source and Nightscout token-based authentication?
Or is that a no-go at the moment?
marafie
@marafie
Jul 31 2017 12:22
@LilDucky nope
if this is true i am blocked how did the installation reach to this stage ?
Tim Street
@tim2000s
Jul 31 2017 12:41
@ericfeibelman You can use a G5 on xDrip+ after a battery replacement and upload the data to Dexcom share from xDrip+, but you can't use the Dexcom app & clarity won't recognise the data past the expiry date. Nightscout, however, will pull the xDrip uploaded data from share.
LilDucky
@LilDucky
Jul 31 2017 12:42
I found it by running ps -aux from another ssh window, saw the git command and then manually ran the git command with http: instead of git:
scharfenort
@scharfenort
Jul 31 2017 12:45

Hi all together, monthes ago, I Setup a Raspberry with openAPS. But with a None loopable Medtronic VEO Pump. I only want to have the Raspberry as a "uploader Station" at home... It is working really very good!!!! Compliment to all involved People!!!

My next question is, if there is a possibility to have a mobile solution in addition. If he is in the kindergarten or at Soccer... Does a mobile solution existing, which will work Close together with the Raspberry solution?

marafie
@marafie
Jul 31 2017 12:48
@LilDucky how can i test it and try it
LilDucky
@LilDucky
Jul 31 2017 12:50
As I said, when the install seems to hang, run ps -aux|grep git from another ssh window. If that task is running then either edit the script or run it manually with http: replacing git:
mkericson
@mkericson
Jul 31 2017 13:02
I'm testing out offline mode for the first time by plugging my G5 receiver straight into OTG w/charger in UART. Do I need to restart the rig when I do this? Then the first indication that it is working will be a temp on the pump- assuming it calls for one?
Diabetic_Dude
@diabetic_dude_twitter
Jul 31 2017 13:49
@mkericson did you ever get a temp ?
marafie
@marafie
Jul 31 2017 13:53
@LilDucky just tested it what you said ,,,, yes it seemed to be blocked
mkericson
@mkericson
Jul 31 2017 14:43
@LilDucky no. reverted back but I didn't get a BT connection, but updated to IOS 10.3.3 this AM as well - so perhaps it lost is BT tether. Trying offline mode again now. I let it run for a while to see if it sends a temp to the pump.
katie disimone
@Kdisimone
Jul 31 2017 14:45
@mkericson are you using a 2017 explorer board. Are you sure you have your OTG cable plugged in in the right direction?
Is your dexcom showing a charging symbol when plugged in?
mkericson
@mkericson
Jul 31 2017 15:09
@Kdisimone Yes. New explorer board v1.2, Yes Dexcom shows charging symbol. Have the host side to the rigs OTG. I'm running good 96 so it likely doesn't need to post a temp yet.
Scott Leibrand
@scottleibrand
Jul 31 2017 15:16
@scharfenort yes, you can build an Explorer board rig as well. They'll play nicely together, taking turns talking to the pump.
@mkericson unless you explicitly disable it, OpenAPS will always set a temp (a neutral one equal to basal if nothing else is needed).
scharfenort
@scharfenort
Jul 31 2017 15:22
@scottleibrand do you mean the riley link, or another solution? Do you have a link for me, which provides some informations(needed Hardware, requierments, how to etc.)...
Scott Leibrand
@scottleibrand
Jul 31 2017 15:23
OpenAPS docs. Surely you used them to set up your first rig?
scharfenort
@scharfenort
Jul 31 2017 15:28
cool, yes, I used it for the Raspi Setup. Bit I forgot that this was also for a mobile RIG. That means, easliy order the edison board and follow the instructions. After that, the Mobile RIG and the Raspi RIG will work together? Great :-D Than I will start the next Level for me :-D
Thanks!
Steve Bernacki
@sbernacki
Jul 31 2017 16:05
Hello there, I built a Edison rig for my wife and she started with it yesterday Medtronic 523 pump (firmware v2.3A 1.1 0B 0B), Nightscout 0.9.2, Dexcom Share, oref0@0.5.2 - autosens on, autotune off. We're definitely seeing the rig picking up comms from the pump (can see when she meal boluses etc), but how can I tell if the rig is actually setting temp basals on the pump?
Dana Lewis
@danamlewis
Jul 31 2017 16:12
@sbernacki look at the pump and see if it sets a temp.
If it's not, number one reason why temps don't get set is you have it in % mode and need to switch temp basal type to u/hr (absolute). If not, paste your logs and we can help you troubleshoot
@dramageek xdrip you'd put as CGM source, token auth is under the advanced features menu in setup script right not
Andy Sharrow
@dramageek
Jul 31 2017 16:17
@danamlewis Correct, that's how I have it set up. But using token-based changes the cron string variable, so the check xdripaps uses fails.
I think. I'm not sure how to test to make sure that's what's going wrong. I did just open an issue with @colinlennon on the xdripaps page, as it seems like that's where the actual problem may be. Otherwise, we may just have to warn not to combine both xdrip and token-based.
Steve Bernacki
@sbernacki
Jul 31 2017 16:21
Thanks @danamlewis , I'll have her check that and get back to the group here. :)
Dana Lewis
@danamlewis
Jul 31 2017 16:23
@dramageek hm, I'd poke @PieterGit and see if he has any ideas on that, too.
@sbernacki :+1: if you've got the blue basal render line turned on, it'll also be visualized that way in addition to showing in the logs. But when you are new to logs, looking at pump is easiest confirmation.
Phil
@drnoname82_twitter
Jul 31 2017 17:56
Is there any way I can force the rig to read ISF from the pump again? We changed ISF in the pump maybe ten hours ago, but the rig still uses the old ISF
Dana Lewis
@danamlewis
Jul 31 2017 17:57
Is that because you have Autotune on, and it's using the Autotune profile? Pump history gets refreshed periodically and should have been pulled many times since then
Phil
@drnoname82_twitter
Jul 31 2017 18:03
Yes, autotune is on, but Autotune only shows the first ISF of the day and even corrects it from 50 to 51 while Autosens corrected the daytime ISF from 40 to 36 causing hypos in the morning.
Dana Lewis
@danamlewis
Jul 31 2017 18:04
Autotune only supports one isf
Dana Lewis
@danamlewis
Jul 31 2017 18:09
If autosens is detecting resistance, there's probably something else going on that's influencing that.
If you want to pull logs from that time frame, you can dig into what it was calculating and why
Scott Leibrand
@scottleibrand
Jul 31 2017 18:16
if you want to "reset" autotune to use the new pump ISF, you'll have to either change the file manually, or move/remove the autotune directory and re-run it.
PieterGit
@PieterGit
Jul 31 2017 18:31
@dramageek @scottleibrand : i think the fix for openaps/oref0#577 is wrong (although the PR fixes an issue). All other openaps / oref0 commands start from the openaps dir, so I think we need to apply constency in that. I didn't notice the problem because I always create the xls with a seperate command. BTW whats the difference between oref0-autotune.py and oref0-autotune.sh. Shouldn't one be removed from the git. I noticed the py-script is not mentioned in the package.json
Andy Sharrow
@dramageek
Jul 31 2017 18:39
@PieterGit IDK about the .py vs .sh. I was just tracking down a problem I had making the .xlsx.
I reopened openaps/oref0#576 that just changes the help text to clarify, then. I'd opened it, but closed it when I figured out how to do #577.
PieterGit
@PieterGit
Jul 31 2017 18:40
See colinlennon/xDripAPS#2 for my comments on tokenbased authentication and xdripaps
Scott Leibrand
@scottleibrand
Jul 31 2017 18:46
@PieterGit afaik the autotune xlsx stuff has always been broken, but most people don't try to use it, so happy to merge (and/or revert) whatever y'all think is best there.
PieterGit
@PieterGit
Jul 31 2017 18:47
let's apply the rule that all openaps/oref0 commands need to be executed from the openaps dir. I can't do any git changes the coming week, but will make the changes later (if @dramageek wants to give it a try, please do)
Scott Leibrand
@scottleibrand
Jul 31 2017 18:48
:+1:
Andy Sharrow
@dramageek
Jul 31 2017 18:49
In that case, apply #576 and discard #577.
Sorry for the confusion!
Dana Lewis
@danamlewis
Jul 31 2017 18:49
no worries, thanks for jumping in with PRs @dramageek
PieterGit
@PieterGit
Jul 31 2017 18:53
@scottleibrand : i noticed openaps/oref0#563 is not merged to dev yet. i also fixes some autotune problems
Scott Leibrand
@scottleibrand
Jul 31 2017 19:01
Opened openaps/oref0#578 for that
PieterGit
@PieterGit
Jul 31 2017 19:05
@dramageek : can you be a volunteer for openaps/oref0#578 ? That fixes a autotune error with token based authentication. I have run it for several days without problems.
PieterGit
@PieterGit
Jul 31 2017 19:14
i switched autotune off for now, because of too many lows. Can it be that not entered carbs for hypos/snacks causes higher insuline basal on the next day with autotune? Does autotune also take UAM carbs into account?
Scott Leibrand
@scottleibrand
Jul 31 2017 19:15
autotune does attempt to identify and ignore UAM carb-absorption periods , but I haven't tested that extensively. your autotune-prep log should show which data was attributed to what
Pete Schwamb
@ps2
Jul 31 2017 19:16
What are UAM carbs?
PieterGit
@PieterGit
Jul 31 2017 19:17
UAM = unattended meals
Dana Lewis
@danamlewis
Jul 31 2017 19:17
unannounced meals
Scott Leibrand
@scottleibrand
Jul 31 2017 19:17
sorry, "UAM carbs" is the wrong way to say that: updated
Andy Sharrow
@dramageek
Jul 31 2017 19:18
@PieterGit I'll take a look and see if I can test it somewhere in the next couple days. I'm currently back to using the old API method, and I know I'll need that this weekend, but there might be time in between to test it. Do you need a regular-scheduled autotune run, or could I switch, do a one-of, then switch back?
PieterGit
@PieterGit
Jul 31 2017 19:18
sorry, my bad. should have check the docs :smile:
Phil
@drnoname82_twitter
Jul 31 2017 19:18
My feeling was that very late carbohydrate absorption (noodles being absorbed in the middle of the night) cause ISF to be reduced by autosense and that the reduced ISF then causes lows in the morning hours
Dana Lewis
@danamlewis
Jul 31 2017 19:19
@drnoname82_twitter we don't have to rely on feelings, let's look at logs and see what happened and when, and see what you might want to tweak in future if you want different outcomes :)
PieterGit
@PieterGit
Jul 31 2017 19:20
@dramageek : both are fine. Just like xdripsaps is not compatible with token based authentication. current 0.5.2 autotune is also not compatible with token based authentication. i fixed that with openaps/oref0#578
Scott Leibrand
@scottleibrand
Jul 31 2017 19:21
@drnoname82_twitter in this case, you can specifically see whether a particular late-night rise was used for autotuning basals, ISF, or something else, and determine just how much autotune adjusted things as a result of that day's data.
Phil
@drnoname82_twitter
Jul 31 2017 19:21
I need your help for that ;) autosens-looplog does not give me the answer because I cannot see the times of the deviations taken into account...how do I get that information?
Scott Leibrand
@scottleibrand
Jul 31 2017 19:21
the file you want is in the autotune directory
cd ~/myopenaps/autotune && ls -lart will show something like:
drwxr-xr-x 11 root root   4096 Jul 30 10:12 ..
-rw-r--r--  1 root root   2108 Jul 31 00:05 profile.pump.json
-rw-r--r--  1 root root 240709 Jul 31 00:05 ns-treatments.json
-rw-r--r--  1 root root 137942 Jul 31 00:05 ns-entries.2017-07-30.json
-rw-r--r--  1 root root   1965 Jul 31 00:05 profile.1.2017-07-30.json
-rw-r--r--  1 root root  83959 Jul 31 00:09 autotune.1.2017-07-30.json
-rw-r--r--  1 root root   1963 Jul 31 00:09 newprofile.1.2017-07-30.json
-rw-r--r--  1 root root   1963 Jul 31 00:09 profile.json
drwxr-xr-x  2 root root  53248 Jul 31 00:09 .
-rw-r--r--  1 root root   2052 Jul 31 00:09 autotune_recommendations.log
-rw-r--r--  1 root root  35183 Jul 31 00:09 autotune.2017-07-31-000504.log
then you can use less to look at the contents of that last file (autotune.2017-07-31-000504.log or similar)
(someone should probably PR this in as a "how to drill down if you want to figure out why autotune made the recommendations it did" or something like that)
Phil
@drnoname82_twitter
Jul 31 2017 19:25
Autotune did not change my basals in any significant way yet. I think the too aggressive ISF may be worth examining. How do I find out whether very late carb absorption caused the ISF correction from 40 to 36?
Scott Leibrand
@scottleibrand
Jul 31 2017 19:26
do what I just said, and look for which lines were attributed to ISF. they'll say ISF at the end, like:
0 mealCOB: 0.0 mealCarbs: 0 basalBGI: 6.0 BGI: -4.9 IOB: 0.9 at 00:48:24 dev: -0.39 avgDelta: -5.25 ISF
0 mealCOB: 0.0 mealCarbs: 0 basalBGI: 6.0 BGI: -4.6 IOB: 0.8 at 00:53:24 dev: 0.64 avgDelta: -4.00 ISF
0 mealCOB: 0.0 mealCarbs: 0 basalBGI: 6.0 BGI: -4.3 IOB: 0.7 at 00:58:24 dev: 0.58 avgDelta: -3.75 ISF
0 mealCOB: 0.0 mealCarbs: 0 basalBGI: 5.6 BGI: -4.1 IOB: 0.7 at 01:03:25 dev: -0.67 avgDelta: -4.75 ISF
Phil
@drnoname82_twitter
Jul 31 2017 19:27
Ah ok, so autotune and autosens both use the info contained in that file?
Scott Leibrand
@scottleibrand
Jul 31 2017 19:27
and then just before it shows you the autotune_recommendations.log, it'll have something like p50deviation: -0.92 p50BGI -2.17 p50ratios: 1.377 Old ISF: 55.514 fullNewISF: 76.443 adjustedISF: 63.2215 newISF: 56.285 that shows you what it did to your ISF based on that day's data
no, everything I'm talking about here is for autotune
if you're talking about autosens, I need to switch gears
Phil
@drnoname82_twitter
Jul 31 2017 19:28
Sorry ;)
Scott Leibrand
@scottleibrand
Jul 31 2017 19:28
all that is more for @PieterGit then. :)
can you paste what you have in your autosens-loop.log?
Phil
@drnoname82_twitter
Jul 31 2017 19:40

I hope this is the part of interest:

----+++++---++++--=-------+-+++-+++++++++--++-+++++>>++-----++++>>++++++++>>>>++>>>>>>>>>>>>>>>>>>+++++>> Found 77 deviations since Sat Jul 29 2017 21:47:24 GMT+0200 (CEST) - adding 6 more zero deviations
30% of non-meal deviations <= 0 (target 45%-50%)
Excess insulin resistance detected: Ratio limited from 1.606176470588235 to 1.2
ISF adjusted from 40 to 33

How can I find out which higher-than-expected deviations were taken into account? Without autosense I would have described the past few days as insulin-sensitive (lots of hiking) with some highs due to very late pasta absorption.

Scott Leibrand
@scottleibrand
Jul 31 2017 19:45
the - + and > tell you that, but you have to infer which ones corresponded to which BG data points (or uncomment some debug lines to have it print that)
> means that a given data point was more than the threshold (of 6 mg/dL/5m) and was excluded from the percentiles
I think the ordering is "backwards", with the most recent BG's on the left and the 24h-ago ones on the right
PieterGit
@PieterGit
Jul 31 2017 19:46
@scottleibrand : i'll read your comments and check my autotune files later. i'm out for now. Please merge 578 to dev once approved by somebody else. i'll be less on gitter the coming days, but please tag me if there you think there are questions i can answer, just like dana did today.
Scott Leibrand
@scottleibrand
Jul 31 2017 19:47
:+1:
marafie
@marafie
Jul 31 2017 19:52
For the oref setup ...if I do not use NS and choose No in the setup it would not ask me for the secret API which I need to use in source of CGM .... and once I select the source of CGM as xdrip ...... the xdrip+ app would require http://secretapi@rigs_local_ip:5000/api/v1
Any idea
How can I solve this
I am getting unable to REST API upload :unauthorized url ..... marking record failed
Scott Leibrand
@scottleibrand
Jul 31 2017 19:54
you should just go ahead and use NS?
Andy Sharrow
@dramageek
Jul 31 2017 19:54
@marafie is there a reason you're not using nightscout too?
Scott Leibrand
@scottleibrand
Jul 31 2017 19:54
not sure why we even ask that as a y/n question: everyone should always use NS
Phil
@drnoname82_twitter
Jul 31 2017 20:00
That reverse order in the autosense output would make sense . Lows in the morning hours and quite a few + during the very late pasta absorption. However, when I look at the autotune output for the same time there are virtually no periods attributed to ISF with a positive deviation. Most of the night is (correctly) attributed to UAM.
Could it be that autosense also takes UAM periods into account to determine ISF, thus leading to too low ISF recommendations?
Scott Leibrand
@scottleibrand
Jul 31 2017 20:01
autosens is dumb, and doesn't know anything about COB or UAM. it just has a hard cutoff (something like 6 mg/dL/5m IIRC) and ignores any deviations above that as likely caused by carb absorption or something else that wouldn't warrant counting toward resistance
so if you have meal periods where you're so sensitive that your deviations don't even get that high, then it can misattribute meals as potentially indicating resistance
see openaps/oref0#430
Andy Sharrow
@dramageek
Jul 31 2017 20:05
@marafie Even if you're planning to use xDrip for your CGM source, you still need some way to see what the rig is doing beyond staring at the logs. It's a LOT quicker and easier to check the status through Nightscout than filtering through the logs to find what you're looking for.
Phil
@drnoname82_twitter
Jul 31 2017 20:10
That makes sense. Deviations during the UAM periods (which autotune recognizes with amazing precision) are mostly between 1.5 and 4. So I will limit autosens_max to 1.05 as a first measure.
Is there a reason why autosens does not make use of the autotune attribution to ISF/basal/CSF? Would it maybe make sense to lower the hard cutoff?
Or is that just because autosens is older than autotune?
marafie
@marafie
Jul 31 2017 20:11
@scottleibrand once I travel its hard to get SIM card for each device I have
Scott Leibrand
@scottleibrand
Jul 31 2017 20:11
installing oref0 to use NS doesn't mean you always have to use it
but when you happen to be on wifi, it should
marafie
@marafie
Jul 31 2017 20:12
I have been running offline for 2 months
And I am happy
Scott Leibrand
@scottleibrand
Jul 31 2017 20:12
@drnoname82_twitter exactly: autosens was done before autotune, so we haven't back-ported that logic to it yet.
marafie
@marafie
Jul 31 2017 20:12
Today I installed oref1
Scott Leibrand
@scottleibrand
Jul 31 2017 20:12
NS is really useful for monitoring what oref1 is doing: I highly recommend it
marafie
@marafie
Jul 31 2017 20:12
Can't get the phone uploader to upload to the. Reg
I am too with using it
When I travel its another issue
Scott Leibrand
@scottleibrand
Jul 31 2017 20:14
you can have both xdripaps and NS configured in your xdrip app and it'll upload to both
marafie
@marafie
Jul 31 2017 20:14
Hoping when we have openaps app this would be solved I think
Dana Lewis
@danamlewis
Jul 31 2017 20:14
we'll likely still recommend NS
Scott Leibrand
@scottleibrand
Jul 31 2017 20:15
NS gives you a much better historical view of what happened at X point in the last 48h
Dana Lewis
@danamlewis
Jul 31 2017 20:15
but in the meantime: the answer to your question of how to do that, is to set up NS for now. it won't hurt when you don't have internet to the rig.
so instead of creating a custom, complicated work around, it's easiest to just set up NS
marafie
@marafie
Jul 31 2017 20:15
I have it enabled but I disabled it
Dana Lewis
@danamlewis
Jul 31 2017 20:21
k, then say Y in setup script to NS and then you can provide your API secret there
Phil
@drnoname82_twitter
Jul 31 2017 20:23
@scottleibrand Thank you for your explanations!
Dana Lewis
@danamlewis
Jul 31 2017 20:23
@drnoname82_twitter it would be great to PR a section to "understanding autotune" if that was helpful! (or ditto for autosens)
Phil
@drnoname82_twitter
Jul 31 2017 20:24
I will write a PR for the autosens section
Dana Lewis
@danamlewis
Jul 31 2017 20:24
:+1:
Phil
@drnoname82_twitter
Jul 31 2017 21:08
I just read that limiting autosens-max also limits autotune. Is there a way to disable autosens without having to do the entire setup again?
Dana Lewis
@danamlewis
Jul 31 2017 21:09
if you want to not use it, edit your runagain file to not include it in the enable list, then do bash run again to run it.
Phil
@drnoname82_twitter
Jul 31 2017 21:10
Thanks!
Dana Lewis
@danamlewis
Jul 31 2017 21:10
:+1:
Paul Andrel
@stavlor
Jul 31 2017 21:25
gotta love those branch names
Dana Lewis
@danamlewis
Jul 31 2017 21:26
docs, or code, or both? :D
Paul Andrel
@stavlor
Jul 31 2017 21:27
docs for the most part
Scott Leibrand
@scottleibrand
Jul 31 2017 21:29
patch-1 FTW!
Dana Lewis
@danamlewis
Jul 31 2017 21:29
@scottleibrand probably this one is why :D openaps/docs#988 #literalnamingofbranches
Paul Andrel
@stavlor
Jul 31 2017 21:29
@scottleibrand lol
Yeah githubs lovely pr tool makes interesting branch names sometimes, I was too lazy to bother cloning it then creating a branch and then doing it
Its better than my work project stuff all those things are branch named under ticket numbers aka you need a tool to tell you what it is
aka wip/INC992 crazy cause you need to pull up jira just to find out what the hell that branch was for
Dana Lewis
@danamlewis
Jul 31 2017 21:32
yeesh
Ljuslykta
@Ljuslykta
Jul 31 2017 21:35
I'm pretty new in the loopinh buissness and have had it running for a couple of days now. I'm getting some highs and lows like a sine wave, so I feel that sone tweaking needs to be done. Is there any document on the site that describes the
Scott Leibrand
@scottleibrand
Jul 31 2017 21:36
@Ljuslykta I would look at running autotune
Dana Lewis
@danamlewis
Jul 31 2017 21:36
I would read http://openaps.readthedocs.io/en/latest/docs/While%20You%20Wait%20For%20Gear/Understand-determine-basal.html (to understand what goes into the math and the logs), plus http://openaps.readthedocs.io/en/latest/docs/While%20You%20Wait%20For%20Gear/preferences-and-safety-settings.html to understand the different settings and what you might want to tweak and examine the impacts of.
Ljuslykta
@Ljuslykta
Jul 31 2017 21:36
Approach of iterativley changeing
The values. Sorry for the chopped messages, my phone is acting up
I'll start there
Thanks
Dana Lewis
@danamlewis
Jul 31 2017 21:37
no worries. check out the first link I posted for some background reading. :+1: and let us know if you have any questions, or post an example scenario and we can talk through it if you are not sure why it's doing a particular thing.
Paul Andrel
@stavlor
Jul 31 2017 21:45
I find its usually relatively simple to get most of the reasoning behind why it does something IE: ```
BG: 208, COB: 0, Dev: 73, BGI: -15.16, ISF: 35, Target: 110; Eventual BG 161 &gt;= 140, adj. req. rate: 3.57 to maxSafeBasal: 3.4499999999999997, temp 3.4499999999999997 &gt;~ req 3.4499999999999997U/hr IOB: 3.43U, Basal IOB 0.51U
Ljuslykta
@Ljuslykta
Jul 31 2017 21:46
There's one other thing. In Nightscout, I can easily see the battery levels of the devices, and a graph over the past basal values. Is there a way to see those things in Xdrip+?
Paul Andrel
@stavlor
Jul 31 2017 21:48
Hmm, don't think so in xdrip but if you compile/Build AndroidAPS it does have options for seeing most of that straight from NSClient
Can get nice little Basal/IOB/Dev Graphs too if configured correctly
Scott Leibrand
@scottleibrand
Jul 31 2017 21:49
@stavlor even better than branch names like patch-2 are PR names like Update edison.md;-)
Paul Andrel
@stavlor
Jul 31 2017 21:49
:+1:
Dana Lewis
@danamlewis
Jul 31 2017 21:49
@stavlor want to add to this section of the docs about that? http://openaps.readthedocs.io/en/latest/docs/Customize-Iterate/offline-looping-and-monitoring.html#xdripaps-for-offline-bgs-and-looping-for-android-users not sure if references AAPS for also good vis for offline OpenAPS-rig based looping.
Paul Andrel
@stavlor
Jul 31 2017 21:50
Certainly can look at adding something when i get home time to go play in traffic
Dana Lewis
@danamlewis
Jul 31 2017 21:50
:smile: :+1:
Marco
@CaptainBalou
Jul 31 2017 22:35
@sulkaharo I am on your new curves branch for testing the behaviour. Today at 11am my loop stopped working completely. I now investigated why and found out that autosens.log is growing +3MB a Minute. So my filesystem was full. The message which is written there is a million times "Pump DIA must be set to 5 hours or more with the new curves, please adjust your pump. Defaulting to 5 hour DIA.“. Good hint to read this in the log but would be sufficient to have it once per minute or so. ;-) I now corrected that in my pump and the message in the log is gone as well.
Paul Andrel
@stavlor
Jul 31 2017 23:31
Lol so you had 3MB/min of that nice
live4sw
@live4sw
Jul 31 2017 23:44
Quick question - I got a new phone. Do I need to run oref0 setup again with the bluetooth address flag specifying the address of my new phone? Or can I just follow the bluetoothctl steps to set up the new phone?
Dana Lewis
@danamlewis
Jul 31 2017 23:52
the latter, PLUS editing it in your run again and either running again, or also changing in your cron.
(so you don't forget it for the next time you run the script)